A framework for the establishment and utilization of national geographic information systems

UNITED NATIONS

ECONOMIC COMMISSION FOR AFRICA

Geographic Information Management

the nature of resource and environmental information, data sources and their organization

A FRAMEWORK FOR THE ESTABLISHMENT AND UTILIZATION OF NATIONAL GEOGRAPHIC

INFORMATION SYSTEMS

December 1993 - Revision June 1994

Provisional document

ECA/NRD/CRSU/93-7 Distn LIMITED

Original: English

Geographic Information Management

the nature of resource and environmental information, data sources and their organization

A FRAMEWORK FOR THE ESTABLISHMENT AND UTILIZATION OF NATIONAL GEOGRAPHIC

INFORMATION SYSTEMS

UNITED NATIONS

ECONOMIC COMMISSION FOR AFRICA December 1993 - Revision June 1994

Provisional document

CONTENT

Introduction Scope Justification

Section One: Basic conditions and guidelines

At national level

Understanding of modern information technologies 1

Appreciation for base-line spatial data and information 2

The will and commitment to change 2

A plan of Action 2

National steering body on Geomatics 3

At institutional level

Variety of databases and GIS 5

Building and maintaining databases 5

Organizational and cultural conditions 5

Training 6

Database Administrator 7

Database design and conversion 7

Pilot projects 9

The role of the private sector vs. the public sector 10

Typical components and steps of an institutional

GIS implementation process 11

Section Two: Geographic information management

The nature of geographic information 13

Cataloguing of information 13

Data representation 14

Structures of geographic data storage 15

Major data sources 16

Further characteristics of geographic data sources 18

Data organization 20

The nature and importance of GIS 22

GIS functionalities 23

Creating the database 24

Data integrity 27

Geometric integration 28

Central Database vs. Distributed Database 29

Other aspects of integrity 32

Appendix 1: From the real world to computer datafiles 35

Geographic data files 35

GIS components 41

GIS functionalities 45

Appendix 2: Chapter 40 of Agenda 21: "Information for

decision-making" 47

Annex 1: Accuracy and interpretability of satellite images 54

Bibliography 55

INTRODUCTION

Scope

This document attempts two things: (a) to provide those who are responsible of the management of the land and the protection of the environment (usually non-specialists on computer spatial data manipulation) with a simplified vista on the capabilities and potentials of Geographic Information Systems (GIS) and to familiarize them with the means the computer uses to represent and manipulate the real world, and (b) to provide a set of general conditions and rules, exclusive of a particular sector or user's view, on how to proceed for the establishment of a national GIS composed by a network of databases and individual institutional GIS.

Although implementing GIS entails, in practice, highly specialized computer hardware and software, and expertise, this document avoids technicalities normally of interest to the specialist but not to the decision-maker who must take decisions concerning the implementation and

management of a GIS.

The first section outlines a number of basic conditions at government and institutional level that are thought critical to the establishment of a national spatial information infrastructure. It lists the major steps towards the establishment of an institutional GIS, within the scope of a nationwide information system. Appendix 1 supplements this section by giving an outline of how geographic data is stored, structured and handled by the system.

The second section focuses on general spatial information management and includes issues such as the nature and importance of geographic information, and its accessibility and manipulation.

Appendix 2: Chapter 40 of Agenda 21: "Information for decision-making", should be a helpful

addition.

Justification for these guidelines

GIS is becoming the new creed for those who deal, in one way or another, with land resource

development, environment protection and other activities related to land occurrences. This phenomena

is, to a great extent, a result of the power of these systems to cope with increasing amounts of

varying spatial information, and its ability and flexibility to process and display data according to the

different purposes and planning scenarios. It can also be misleading to those who want a GIS to solve

all their problems. Some planners may feel that GIS must be introduced at any cost and without delay

in their operations, without fully understanding all the implications of establishing a GIS, much less

the kind of GIS that is suitable to them. Surrounded by an overwhelming amount of GIS-promoted

software available, ranging from the complex and expensive to the simple and cheap, many organizations are tempted to purchase one particular software and begin to collecting digital data, often on ad-hoc basis, in the belief that they are establishing the GIS they need. Within a region, a

country, a municipality or even within an institution, a proliferation of ill-suited GISs will most likely

result in nothing more than a lamentable waste of resources.

On the other hand, many may be reluctant to enter GIS for fear of getting involved in something that is too technical, too costly, implies too much work and is not worth it.

When these two attitudes -the positive-but-naive and the negative- persist, the high front-end costs coupled with confusion and frustration will by far outweigh any gains, and what initially were dreams and expectations turn into nightmares. The results, rather than beneficial, will prove damaging and

may delay or destroy any further attempts to use the technology.

In order to avert these situations, GIS developers, producers, scholars and users, usually from the

private sector, have generated various models providing rules and guidelines on how to construct a

GIS. However, they are part of the scientific literature and are "promulgated" in specialized

publications, dispersed among or entwined with technical articles of all sort, and do not the reach the

normal planner. This document compiles, reconciles and integrates many schemes proposed by

renowned scientist and specialists as well as published technical material.

SECTION ONE

FRAMEWORK OF BASIC CONDITIONS AND GUIDELINES FOR THE ESTABLISHMENT OF A NATIONAL RESOURCE AND ENVIRONMENTAL

GEOGRAPHIC INFORMATION SYSTEM.

There is no general model that suits every country for the establishment of a national geographic information system, as there is no general set of procedures or steps to implement such a system.

These is a large range of conceptual models and implementing procedures among which to choose to better satisfy the needs and peculiarities of each individual country. However, a number of basic enabling conditions and guidelines are identifiable as being crucial to the success of building up a useful geographic information infrastructure in any country. These exist at both national level and institutional level. An attempt is made here to provide those that are thought of major relevance.

At national level.

Understanding of modern information technologies and their potentials by those who have the keys to development

It is essential a full understanding of the potential of modern information technologies, in particular geomatics, and of the need to use then in the inventorying, assessment and management of natural resources and the environment, and in setting up and steering sustainable development paths that will solve socio-economic problems of African countries, all urgent and overwhelming.

It must be realized that the possession of a reliable and efficient spatial information infrastructure

is as important as other national infrastructures upon which the governments concentrate their

efforts, such as health, education, transport, energy, etc. Making a parallel, what expectations could have an international airline without possessing a reservation database server or that is not linked to the global reservation network?. How could a modern bank presently operate without possessing a client/account information system networked with its branches and subsidiaries?.

Such an understanding must be done at the highest levels among planners, decision and policy makers, in the government (heads of state, ministers, etc.), in the political scenario (chiefs of political parties, senators, in some cases governors and majors, local community leaders, etc.), in the production and industry sectors, etc.

A framework for the establishment of national GIS ECA/NRl^/CRSU/93-7 Section One: BASIC CONDITIONS AND GUIDELINES Preliminary document - Revision June 1994

The appreciation for base-tine spatial data and information.

Base-line data and information, in particular topographic and other land information maps, is normally the least appreciated tool. National mapping programmes, which in the past were at the forefront of development, are nowadays given little preeminence in national development programmes. As a consequence, the cartographic coverage of the majority of African countries, with some notable exceptions, is deficient and outmoded, in some areas nonexistent or the mapping is only planimetric, and in others little attention was given to geodetic control. Although valuable efforts have been made by some leading countries, the situation for many of them remains unchanged and may worsen as time elapses. The governments must grant to this activity a high priority, allocating the resources that it needs to complete, improve and maintain their national cartographic coverage. Otherwise, their efforts to build any national geoinformation infrastructure will be futile, and whatever is done will lead only to a lamentable waste of resources and an anthology of frustrations. Simply, a building without foundations cannot be erected.

The will and commitment to change.

Understanding potentials and realizing needs is not enough. It is necessary to have the commitment to change, by conviction. Only then we can expect to practically materialize this change.

This commitment has to be a long-term commitment. It has to last at least until the development of the system is well in place, and its benefits appear. Then the users (among which the government will be the major one) will be the driving force to maintain and continuously expand the system that may

become self-sustainable.

The materialization of the will to change: A Plan of Action (or the implementation ofAgenda 21).

A plan of action, issued at the highest level, that would lead to a national programme for the creation and management of a geographic information system, would be the first step to materialize the will to change. Such a Plan of Action would be a great leap in the implementation of chapter 40 of Agenda 21. A copy of this chapter appears as Appendix 2. Such a plan of action can (should) initially be simple and should be inscribed within a broader Plan (Agenda 21). It would, at further stages and as results of its own implementation, be amended and improved, going in deeper detail, changing strategies and procedures, adding new components, fixing new responsibilities, etc.

This initial plan of action would:

(i) provide principles on flow, access and supply of geographic information.

(ii) define the goals at short, medium and long-term.

(iii) identify the sectors that would be addressed by the system: principal, secondary, tertiary.

(iv) identify the actors including those of the private sector; national and sub-national agencies, research and educational institutions, agricultural and livestock associations,

-2-

A framework for the establishment of national GIS ECA/NRD/CRSU/93-7 Section One: BASIC CONDITIONS AND GUIDELINES Preliminary document - Revision June 1994

scientific and professional associations, etc.

(v) create a national steering body, constituted by a core set of actors, including the private sector, with the task of defining the nature, characteristics of the system as the methodology,

procedures and time-frames of implementation. When necessary, any other members can be

coopted by the steering body. This body would be accountable to the highest levels of government, to which it would report regularly at specific intervals.

(vi) Define provisional budgets and budget lines.

A National steering body on geomatics:

It can be an ad-hoc Committee or a permanent Council. Sometimes, a leading Institution may be selected to coordinate the work of the Committee or Council. Its work would mainly be conducted by working groups, whose initial functions would be:

(i) Identify the spatial datasets that are required for each sector or aspect of development. This

exercise will necessarily comprise the identification of data users, data needs and data sources,

required data accuracy, as well as the selection and prioritization of scales.

In general, the following datasets have been accepted as an appropriate basic set for national planning.

Fundamental (a) Topographic map: geodetic control

data sets: elevation (DTM)

water areas, drainage and shorelines terrestrial communication lines cultural elements

general vegetation: forests, rangeland,

cultivated areas and pastures,

principal administrative boundaries

Other data sets (b) baseline satellite image

(c) cadastre, land tenure, detailed administrative boundaries.

(d) geology/mining

(e) land use/land cover

(f) soils and land vocation (g) energy

(h) restrictive sites (public lands, special tenures)

(i) climate

(j) fauna species distribution

It is very important to bear in mind that the topographic map is the basis upon which the other datasets are geometrically fit. Hence, it must have the highest priority.

-3-

A frameworkfor the establishment of national GIS EX^A!^C?S

Section One: BASIC CONDITIONS AND GUIDELINES Preliminary document - Revision June

(ii) Compile a classified catalogue of all existing data and information that is deemed relevant, assessing the attributes of each piece of information. The committee will also be responsible for assuring that such metadatabase is properly maintained.

(iii) Identify datagaps, and provide concrete recommendations for new data collection.

(iv) Select the model of the national geographic information system (central database, distributed database, combined central and distributed) and, accordingly, how the different datasets are organized.

(v) Compile an inventory of all existing hardware and software, within participating parties, qualifying its relevance to the project: operating platforms, capacity, compatibility, risk of out- datedness, flexibility, facility of maintenance, possibility to expand, etc.

(vi) Analyze the current mandates of the different participating institutions. The Committee would agree on new limits to the mandates for the purposes of the establishment of the national geographic information system, in order to avoid the normal conflicts due to overlapping and duplication, or to fill gaps, that inevitably will always be found. Precise responsibilities and roles of each institution would be clearly defined within the system, so that every actor knows what to do and what to expect

without ambiguities.

(vii) In the light of the above, the Committee would propose or set up the rules, as applicable, concerning legal and technical aspects of mandates, proprietary rights, security and confidentiality, flow of information, access and supply, pricing of the information, data quality standards and standards for data collection, data up-dating, data conversion, integrity and integration of datasets

(geometric, datatransfer, etc.)

(viii) At least a written master agreement is necessary to protect the interest of all parties to the extent possible, where the issues in (iv) and (v) above are clearly defined.

(ix) Provide advice on and revise the datamodels and datastructures proposed by each institution for the datasets of which it is responsible for producing and maintaining, verifying that they satisfy the needs of all the development sectors that will use those datasets.

(x) Revise the necessities and procurement of new equipment and software, assuring that there is full compatibility with the equipment and software of the other participants of the systems, and that appropriate existing equipment and software has maximum, but reasonable, possible utilization.

(xi) Design and coordinate the execution of pilot studies and pilot projects, where tests of small integrated systems would be carried out as a means of obtaining practical experience with the technology and its possibilities, and test system conditionalities such as networking and communication among GIS sub-systems, standards, database export-import, data integrity, data flow,

etc.

(xii) Propose, set up and phase implementation plans.

-4-

A framework for the establishment of national GIS ECA/NRI#CRSU/93-7 Section One: BASIC CONDITIONS AND GUIDEUNES Preliminary document - Revision June 1994

At institutional level

Variety of databases and GIS within the system:

It is clear than the datamodels and GIS type and characteristics, as well as the strategies of

implementation within the individual institutions will depend on each institution concerned, and will vary from one to the other. Among other factors, they will have to take into account and will be conditioned by: (a) whether it is a source producer, a user, or a combination of the two, (b) the

type and nature of the data, (c) the volume and complexity of the data, (d) the analysis required

on the data to satisfy the needs of the users, (d) the physical, financial and human resources

available, (d) the institutional structures, (e) etc.

So far, it would be unrealistic to conceive uniform individual systems within the general national geographic information system network. Each institution will be responsible for selecting the GIS

it will use, but the choice must not be isolated from the rest of the system. What is important is the conditionally that they communicate freely and that the databases are integrated together.

Building up and maintenance of the databases:

In a distributed system, with is being adopted more and more widely, the producer (owner) of the

data and information is also their custodian. In such context, he defines his own datatamodel and data structure, but will assure that they satisfy the needs of the users within the national information system, and in such a way as to facilitate the up-date and expansion of the content.

Data conversion, data maintenance and data up-dating will be the responsibility of the producer.

He will also be responsible of making his database fit to the geometry of the fundamental

topographic database. In an ideal case, common elements from the two databases would be adjusted to share the same set of primitives, assuring full integrity.

Hence, the conversion of the fundamental database should be done with priority, as the other datasets are just layers to it.

At any rate, although it implies a greater rate of investment, the conversion of data should be

done with great care and within the shortest possible period. The reason of preferring a short

conversion period is that the effects and benefits will not be obtained until a complete database exists for at least one application theme covering a comprehensive geographic area.

Organizational and cultural conditions:

The introduction of GIS technologies leads to changes not only in existing routines for information exchanges between and within national authorities and agencies, but also entails

changes in old-time conceptions on the nature itself of spatial information, conceptions considered immutable in the minds of those who have been producing and managing that information. These changes imply organizational changes within the institutions and also challenges the cultural attitude of the staff, at all levels, as the new technology threatens their system of values. A certain

amount of the personnel will oppose to them, either deliberately or not.

-5-

A framework for the establishment of national GIS ECA/NRV/CRSU/93-7 Section One: BASIC CONDITIONS AND GUIDELINES Preliminary document - Revision June 1994

Institutions switching to GIS must be aware, from the very beginning, that these organizational and cultural problems are more intricate to solve than technical ones, and at the same time are crucial to the degree of success achieved. A great deal of time and attention has to be devoted these matters.

Clear and open information on the new methods/goals of the organization to all concerned staff (operational, supervisory and managerial) should be a priority, where an internal convincing marketing of the new working procedures, new products and overall expectations would be earnestly conducted.

Awareness and education through periodic seminars and workshops must be organized, encouraging and entertaining open discussions and creating a sense of solidarity and interdependence, whereby the units of the organization would feel genuinely responsible for developing and maintaining the new functions. If an important part of the organization does not accept them, then it should be removed

from any role/activity related to the new system.

Training:

Qualified staff is a sinequanom in geomatics. The success or failure of individual GISes will directly depend on the availability of competent staff, who can understand the processes behind the technology. Adequate and continuous training at all levels is required: (i) Operators for data conversion (digitizing, editing), up-dating and display. Sometimes experienced but enthusiastic cartographers and draughtmen can be successfully (and easily) trained, as they understand, better than others, the mechanics of cartographic data representation; (ii) Supervisors: They must be individuals with a full understanding of the mandates of the organization and of the processes utilized to carry out those mandates, in particular concerning the generation and the purpose of the institutional data that will be stored and used by the system. These individuals must also possess a thorough understanding of the system's datamodel and datastructure, with good expertise in the manipulation of the GIS dataconversion modules, and must understand how the database is to be queried; (iii) GIS applications staff. These are professionals (engineers, surveyors, foresters, soil scientists, environmentalists, urban/rural planners, etc.) specialized in spatial data analysis with GIS, with an intense and dedicated training in the use and characteristics of the particular GIS package being implemented by the organization. Data analysis and manipulation, modelling, planning scenarios, etc., would be the responsibility of this group. It would also be responsible in assuring that the GIS meets the goals that have been set up and that the (growing) users' needs are met; (iv) Institution managers:

A good understanding of the components and of hardware and software, database models and structure, functionalities of the GIS and how the system can be queried and analyzed, is essential, and can be attained through the organization of seminars and workshops; (v) Decision-makers: who are the real end-users of the system. Training would focus on knowing the content and structure of the database and in the use of the GIS modules to manipulate and analyze the database. As a rule, they are assisted in their task by the applications staff.

The training scheme will depend on the organization's structure and its resources, on the type and complexity of the data to be converted and used, as well as in the schedule and time-frame for the system's implementation. Some of the staff may be identified within the organization and some will have to be recruited. Some training can be done on-the-job (operators, supervisors) by the system's vendor during the initial installation phase and pilot projects, or, if available, by contracting with some experienced group in the country. If the data conversion phase is done by an external firm, provisions can be made whereby this process is utilized to provide the required on-the-job training

-6-

A frameworkfor the establishment of national GIS ECA/NRD/CRSU/93-7 Section One: BASIC CONDITIONS AND GUIDELINES Preliminary document - Revision June 1994

to operators. A number of professionals will necessarily have to be sent for formal training and education to universities and specialized training centres, within or outside the country. The Regional Centre for Training in Aerospace Surveys (RECTAS) at Ile-Ife, Nigeria, established under the auspices of the Economic Commission for Africa, constitutes an excellent alternative for training and education in Spatial Information Systems in Africa at the three levels of technician, technologist and postgraduate. It has recently initiated the implementation of a new and full fledged course in Geomatics. The Regional Centre for Services in Surveying, Mapping and Remote Sensing (RCSSMRS), also under the aegide of the commission, although it does not offer regular courses as does RECTAS, organizes short and medium-term training courses in GIS to African nationals.

Further, both centres have experience in organizing customized seminars and workshops, which are apt to provide the required know-how for those in (iv) and (v) above.

Whatever scheme is applied, any public organization will quickly realize that rinding and retaining staff with adequate skills may be a serious problem, as it is the adoption of measures and strategies to solve this problem, such as the introduction of special salary scales and effective incentive

strategies.

Database Administrator (DBA):

The person or a group of persons, normally constituted by computer scientist(s) and specialized in GIS design and management, responsible for the overall control of the database. Among other important tasks, he will decide how the data will be defined and stored (conceptual and internal data definition languages), and structured, he will liaise with partners and users ensuring that the data they require is available and accessed, will define authorization checks and validation procedures, will ensure that the data meets the established standards and that the data and information that is generated flow smoothly both internally and externally. Finally, he will organize the system so as to get the performance that is "best for the organization", and will make the necessary adjustments to respond to changes in requirements. This important and complicated task, which until not long ago was a serious ordeal, has nowadays been eased as most of the major GIS packages of the market have incorporated the necessary utilities of the DBA within the system's DBMS, many of which are performed automatically and efficiently in a way that is invisible to the user.

A GIS has been rightfully compared to a car (Konecny, 1993), where the hardware and software supplied by vendors is the car itself, the data is the fuel and the administrator (manager) is the driver, without which the car could not go anywhere (and would serve no purpose).

If funds are available, the recruitment of a specialized consultancy firm to perform the duties of the Database/System Administrator, can be a simple yet satisfactory solution, at least during the entire period of implementation, until the whole system is working successfully and when experts within the organization are highly knowledgeable and familiar with the tasks of the Administrator.

Database design and conversion:

Database design: During this stage, the content of the database, which is conditioned by the requirements of the users' needs, is defined and documented. A datamodel, which is an abstraction

-7-

Aframeworkfor the establishment of national GIS EX^^C?S

Section One: BASIC CONDITIONS AND GUIDELINES Preliminary document - Reviswn June

of the "reality" of a particular application \ is first developed, representing, in a simplified manner, the entities of interest, the attributes that have to be recorded about those entities, and their releshionships. It is then followed by the development of a database dictionary, which classifies, lists and codifies every theme (and sub-themes), object and entity, down to the last identifiable and meaningful element. The database design should include as well descriptions of the specifications and standards, sources of data, and of the processes for data input and conversion, updates, maintenance

and archival.

At the final stage of the database design, the actual structure of the database is developed and documented against the software platform that, by then, must have necessarily been already selected.

However, the original concept of the database should not be confined or restrained to a particular hardware and software, as the lifetime and cost of the database surpasses by far the lifetime and cost of hardware and software. In this regard, data independence is a major objective of database systems.

Another important conditionality is that the database design takes into account interfaces and communication with existing and planned computerized datasets, as it is the case within a national geographic information system, more so when the datasets are spread across a network of distributed

datasets.

Database conversion: Several options can be used to populate the database. The process can be done by an external contractor, it can be done internally, or part externally and part internally. Each one possesses advantages and disadvantages depending on the singularities of the organization and the

database.

There are, however, good reasons that favour the first one, that is external conversion, at least for the

bulk of the task:

(i) The database can be very large and therefore, it requires a large number of conversion units (digitizers, scanners, editing stations), all of which are costly and which the organization won't use after the database is completed, ft should be borne in mind that maintaining the database will only need a minimum of units, and that data analysis within the organization can nowadays be performed with inexpensive but powerful desktop PCs and software. A large number of experienced operators is also required, who at a later stage may just sit idle, entailing all the problems related to staff redeployment, laying-off, etc. These operators can, of course, be recruited only to perform the task of data conversion, but again they have to be adequately trained and once the process is over, the institution will lose the investment made. Finally, if the organization can not invest in an appropriate number of conversion units or cannot find and train sufficient operators, then the time to populate the database may increase substantially, postponing results (and benefits), and augmenting the risks of failure.

1 Nevertheless, the database system should be independent from the application. In fact, different applications will need different views of the same data, in particular within the concept of a national distributed geographic information

systems, where the different datasets will be use to satisfy the needs of different sectors.

-8-

A frameworkfor the establishment of national GIS ECA/NRD/CRSU/93-7 Section One: BASIC CONDITIONS AND GUIDELINES Preliminary document - Revision June 1994

(ii) The process of data conversion is new to the organization. If not enough care is taken and

quality control is deficient, too many errors are bound to be committed affecting the quality of

the database, which sooner or later will be detected and will have to be corrected, either via re- digitizing and editing and re-editing, large portions of the database. In addition to unexpected

delays, internal and external criticism and pressures, compounded with lack of confidence and

frustration, may show very damaging to the implementation of the GIS.

(iii) Data conversion, as other routine tasks, are often carried out more efficiently by the private

sector than by the public sector (see below the role of the private sector).

At any rate, a close monitoring of the data conversion process must be done, assessing the quality of the quality of results, whereby the data converted is accepted or rejected following a clear set of rules, standards and specifications. The responsibility for such control can be assigned to a unit of the organization, if the required know-how is available in house, or can be entrusted to a specialized external firm.

Pilot projects.

Pilot projects are essential elements of any GIS implementation plan. These pilot projects are necessary to test that all the components and functionalities of the system meet its objectives.

The realization of a pilot project consists of the data conversion of a small geographic area of the dataset, which is loaded into the hardware/software selected. Data content and structure, data storage

and access, data analysis and queries are verified against the original specifications. The pilot project

will also test data conversion and acceptance procedures.

Within the context of the national geographic information system, it is clear that an integrated pilot project, comprising data from distinct datasets from different sources, must be designed and conducted, to test, inter alia, data export/import and integration, compatibility among the individual systems of the network, and pilot applications.

Benchmarks should also be designed and carried out before the acquisition of the particular equipment that the svstem(s) will use is decided. The primary role of a benchmark is to provide an

unbiased mechanism to measure the suitability and efficiency of a supplier's proposed GIS hardware-

software solution within the context of the institutions' requirements and environment. The benchmark content should closely replicate the user's application and dataset characteristics, but should be concise and focussed on key aspects that can quickly exposed any potential inadequacies.

-9-

A frameworkfor the establishment of national GIS ECA/NRD/CRSU/93-7 Section One; BASIC CONDITIONS AND GUIDEUNES Preliminary document - Revision June 1994

The role of the private sector vs. the public sector.

Many tasks are cheaper and more rapidly performed by the private sector than by the public sector.

This fact, until recently considered an anathema by public officials at all levels, in both developed and developing countries, is now being slowly recognized and accepted. With notable exceptions, major public institutions have traditionally, by their nature, a strong inertia with stiff hierarchical structures, inflexible norms and rigid procedures, coupled with a "civil servant mentality", all of which make them difficult to perform fast and efficiently. "Social service", and not rentability, is the driving force. The sense of urgency is replaced by a sense of "gravity", where in order to do things properly, these must be done carefully at a determined pace that can not be accelerated. Taking risks is something that is totally out of question. When, in a given case, a decision is made to move faster,

the organizational inertia will apply the brakes.

The private sector, on its side, is not limited by traditional cultures or procedures. It is therefore more flexible, and not only can adapt easily to changes, but these are welcome. The driving force is cost- effectiveness, and private firms will swiftly test and incorporate innovative technologies and methodologies to speed up production rates while at the same time lowering costs, taking any necessary risks. The staff, in expertise and number, is also elastic and is easily accommodated to fit the needs of the firm. The danger lies on the "responsibility" of the firm, as quality may be sacrificed

on behalf of speed and cost.

To a greater or lesser extent, many States still consider that the private sector: (a) has no business in national mapping or in the collection and management of other types of spatial resource information, or: (b) are competitive bodies and that such competition should not be encouraged or allowed, or; (c)

both of the above.

Such attitudes can not be any longer sustained in a modern market economy, where complementarity and partnership between the two sectors are main conditions for development. The private sector can lend/transfer to the national institutions the expertise and know how they may be lacking, and add to and complement their production capacities, enabling them to absorb new technologies as well as to cope with increasing demands for products and services. On its side, public entities, by contracting part of their activities with private firms, will encourage the development of this sector, not only

creating new jobs and opportunities, but will also ensure the acquisition of appropriate endogenous technical capacities within the country.

Successful examples of such partnership are not difficult to find as it is the case in many countries with cadastre systems regulated and controlled by the State whereas the surveys are carried out by

private surveyors.

-10-

A framework for the establishment of national GIS ECA/NRP/CRSU/93-7 Section One: BASIC CONDITIONS AND GUIDEUNES Preliminary document - Revision June 1994

Typical components and steps of an institutional GIS implementation process

" INITIATIVE TO INVEST IN A GIS

1

* EDUCATION

Orientation to national planners, decision-makers and staff through seminars, workshops, formal and

informal meetings.

• understanding GIS

• relational databases, topology, queries

• implications to the agency's role

• performance expectations

• demostrations

I

CONCEPTUAL DESIGN

Establishes the feasibility of GIS, establishes a concept for the system and provides an overall

implementation strategy.

organizational assessment: establishes a starting point: assets and facilities gaps and deficiencies

staff knowledge and motivation outline of responsibilitites and relashioships among all

institutions involved

identification of users' needs and applications identification and analysis of datasources determing datasets, formats, scales and media determing training needs at all levels

feasibility study: estimation of required resources: invest agressively or within normal budget

cost/benfit analysis

study and appraisal by national steering committee

1 APPROVAL

1

• furhter development

• datamodel

• database

• benchmark^)

SYSTEM AND DATABASE DESIGN and streamlining of system concept

* In cooperation and agreement with the national steering body on geomatics

- 11 -

A framework for the establishment of national GIS ECA/NRD/CRSU/93-7 Section One: BASIC CONDITIONS AND GUIDELINES Preliminary document - Revision June 1994

1

IMPLEMENTATION PLAN

Provides a (normally multi-year) programme of tasks for establishing the GIS revision of resources

strategies for: training

selection and design of pilot projects)

procurement and installation of equipment and software data conversion:

database and system administration data up-dating

data analysis and manipulation overall system appraisal phasing

schedule of tasks

1

IMPLEMENTATION

It is the materialization of the implementation plan

* • training programme launched:

• procurement of equipment: bidytender documents

* analysis of responses, benchmark carried out

* selection & award of contract

site preparation installation acceptance test

• specific operator's training completed

• data conversion process: external: conversion purchase process pilot conversion: test data base

requirements met full conversion

internal: recruitment of GIS data conversion consultant pilot conversion

full conversion

• pilot project: test overall system performance

* test data transfer and communication within the network

* test existeni/immediate applications

* develop and test new applications

test GIS management structure

Reassess budget requirements - cost/benefit analysis 1

DATABASE ROUTINE MAINTENANCE AND MANAGEMENT 1

SYSTEM REVIEW AND EXPANSION

* In cooperation and agreement with the national steering body on geomatics

A framework for the establishment of national GIS ECA/NRQ/CRSU/93-7 Section Two: GEOGHAPHIC INFORMATION MANGEMENT Preliminary document - Revision June 1994

SECTION TWO

GEOGRAPHIC INFORMATION MANAGEMENT

1. The nature of geographic information

In our case, the real world is formed by physical objects/ features/elements (what we will call entities) located on the surface of the earth or below it, having homogenous or heterogenous properties and characteristics (what we will call attributes). These entities can be very simple or very complex, and are combined to form other entities. They all have a location within the space that defines bur three dimensional world, they occupy a portion of this space (length/ volume/ surface), although it can be punctual, and maintain a spatial relationship with the other elements (distance, adjacency, intersection, connectivity, overlapping, inclusion, etc). These elements, when considered with their spatial characteristics, are referred as spatial data, georeferenced data or Reozravhic data.

Geographical data are referenced to locations on the earth's surface by using a standard system of coordinates. The coordinate system may be purely local, as in the case of a study of limited area, or it may be that of a national grid or internationally accepted projection such as the Universal Traverse Mercator Coordinate System (UTM).

Geographical data are very often described in terms of well-established geographical "objects", or phenomena. All geographical studies have used phenomenological concepts such as "town", "river",

"soil association", "district" as fundamental building blocks for analyzing and synthesizing complex information. They are often grouped or divided in hierarchically defined taxonomies, for example the

hierarchy of country-province-town-district, or the hierarchy of soil classification systems or of plants and animals. They are also associated in themes and sub-themes (classes and sub-classes) such as

roads and its types, soils and its types, urban blocks, etc.

Data and information that are relevant to resource and environmental management are normally

collected and produced by a large number of local, national and regional sources, both public and

private, in order to fulfill each one's objectives. This leads to a diversity in its existence, reliability,

accuracy, scale, storage, representation, etc, within a given area. Further, in developing countries, it

is not an easy task to know what information is available, who produces what and where it can be accessed. Too frequently, studies are conducted to meet a particular need or user, but the results are not made available or disseminated to other users that may require it. As a result, existing

information, when obtained, will resemble a patchwork with pieces of rich and adequate information,

adjacent to pieces of poor information or lacking it. Rational integrated analysis is then very difficult,

if possible at all.

Cataloguing of Information - The meta-data-base

In order to avoid confusion, losses of time and resources, duplication and triplication of efforts,

- 13-

A frameworkfar the establishment of national GIS ECA/NRfyCRSU/93-7 Section Two: GEOGHAPHIC INFORMATION MANGEMENT Preliminary document - Revision June 1994

misuse of adequate information as well as use of inadequate information, and all the frustration that may come from the scheme just described, it is necessary, as a first and essential step, to undertake a thorough inventory of all existing information, including actual and potential producers of information that is relevant to natural resources, land planning and the environment. Such a catalogue should not be a simple listing, but should include the attributes of each piece of information: The sources (producer, client, depositary, availability), positional accuracy (for spatial data), up-to- datedness (year of production) and other ancillary information such as scale, projection, completeness, source material from which the data was produced, etc, should be normally part of the attributes. If possible, an atlas depicting graphically the inventory of all available information should be constructed.

Data Representation

Land related information is stored and represented by a number of ways. It can be constituted by texts in the form of books, scientific magazines or journals. It can also be constituted by descriptive characteristics containing the geographic location, the extension and the limits of the elements of interest, in graphic form, where the attributes of each element are either in tabular form, or they are incorporated in the graphics in the form of symbols, or as a combination of the two.

real world (Formats of)i

geographic data representation

/ \

i i

non-graphic graphic

(tables, reports) / \

I i

line maps images

(topographic series, (aerial photos, thematic maps, satellite images)

sketches)

The graphic format, in particular maps, has been used since early and simpler times, probably spontaneously, as the best means to represent data on land resources. It still remains unchallenged and unsurpassed by any other format, noting that in recent times analog graphic representation is yielding a long time reign in favour of digital graphic storage. The importance of mapping can never be overemphasized. The great majority of researches of natural resources and other land occurrences require the positioning of features and observations, and plotting them in map. A map is the most rational kind of spatial representation of data of earth resources and phenomena, either static or

dynamic. A map will always yield the quickest and fullest perception of the information, provide the correlation between objects and phenomena, and enable the measurements of spatial-time characteristics. Geologic exploitation, land management, water resource management, cadastre,

-14-

A framework for the establishment of national GIS ECA/NRD/CRSU/93-7 Section Two: GEOGHAPHIC INFORMATION MANGEMENT Preliminary document - Revision June 1994

infrastructure design and implementation, etc, are practically impossible without cartographic support.

Graphic representation goes from simple sketches without scale nor geographic reference, to full fledged cartographic representations complying with the highest mapping specifications of accuracy

and content.

A means that has proved very effective and valuable in the last decades to handle spatial data is the use of aerial photographs. Not only the photographic image serves as the base on which results and phenomena are delimitated and consigned, but through interpretative processes the task of the technician and scientist in procuring raw data is highly facilitated. Although precise measurements of the elements on the photographs are not possible without resorting the sophisticated equipment and highly technological expertise, a fair assessment of distances, surfaces, altitudes, elevation differences

and slopes, can be made with very simple means and procedures. When the resource scientist or

resource manager requires higher accuracy, then the interpreted elements are readily transferred to a precise map by visual correlation or other methods, and finally the thematic map is produced.

However, this requires that a map be available at a suitable scale, which in Africa is not always the

case.

Orthophotographs and orthophotomaps became popular in the seventies and early eighties.

Orthophotos are photographic images, brought to a predefined scale and where the distortions produced by ground relief and by the non orthogonality of the camera axes have been removed. They

conserve all the richness of the photographic image, while at the same time have the standard precision of the map. An orthophoto and a line-map differ in that the in first one the identification of the details are left to the user, while in the latter, the details have been interpreted and symbolized by the map maker. An orthophotomap is an orthophoto or mosaics of Orthophotos, cut to fit the map sheet format, where the map grid and some linear features have been superimposed. Rural cadastre and land use mapping became the most attractive applications of these products in developing countries.

Another product that has come with great force is the satellite image and its corresponding orthophotomap called the space image map or space map. Satellite images are less distorted that photographic images and covers a broader area. There is a broad constellation of observation satellites collect data of the earth and the atmosphere with a wide number of sensors of different spatial, spectral, temporal and radiometric resolution. These constellation will be continuously improved with satellites possessing better resolving power and other characteristics. Satellite images can be obtained and manipulated in both analog form or digital form.

Structures of geographic data storage: analog and digital

Analog form: Until very recently, this "analog" form has been the only way of representation since the very first map was produced and a different "mode" was not even conceived. Thus, practically the totality of the spatial geographic information that has been produced exist in the form of analog maps and documents.

-15-

A framework for the establishment of national GIS

Section Two: GEOGHAPHIC INFORMATION MANGEMENT

ECA/NRD/CRSU/93-7 Preliminary document - Revision June 1994

Structures of data storage

real world graphic representationi

/ \

i analog

1 maps drawn on paper (symbols drawn over

a reference grid

+

legend)

digital1 i

computer data files (coordinates representing points, lines and polygons

+

attributes)

Digital mode: The developments in computer science has led to a different mode of storage and representation. The geographic data is stored in digital form in computer data files which represent the real world.

2. Major Data sources

Lets try to visualize the most important data sources:

Type

GRAPHICS

Maps Topcgraplric series

Thematic maps geological

hydrogeological

rural cadastre, land tenure

vegetation cover;

land use; soils

Scale

1:500 k/1.000 k 1:200 k/ 250 k 1: 50 k/ 100 k

l:250k-1.000k

l:250k-1.000k

1: 50k on remote areas to 1: 5K on semiurban areas 1:200k/250k 1: 50k/100k

Positional Accuracy

150m-300m * 60m - 75m * 15m - 30m *

**

**

relacc: high abs.acc.: high to very low

**

up-datedness/

repetitivity

low (5 - 20 + years)

adequate

adequate to low

medium 5 to 10 years

low,

natcoverage is normally partial

Supplier

National/Federal Mappin g Institutions

Geol. Dpts. Min. of Mines and Energy, Oil companies

Geol., Nat. Resources and Energy Dpts.

Cadastre, Land Tenure Depts.

Forest, Nat. Res., Agriculture authorities.

Research institutes,

-16-

A framework for the establishment of national GIS

Section Two: GEOGHAPHIC INFORMATION MANGEMENT

ECA/NRfyCRSU/93-7 Preliminary document - Revision June 1994

Images

oceanographic &

hydrographic

rainfall &

veget. index other (census, population, nat.

parks, livestock, fishery, etc

1:500k/1000k 1:200k/ 250k 1: 50k/ 100k

l:4.000k-10.000k

variable

* The national topographic series offer the accuracy is not normally surpassed by other

variable

**

**

very low, partial

coverage

very high -weekly- low,

some produced on ad-hoc basis

Oceanog. &

hydrographic agencies, mapping institutions, Navy forces Meteorological authorities National, federal, municipal authorities, research instit., IGOs, NGOs

highest positional accuracy attainable at the scale of the map. Tills analog maps.

** The uncertainty of boundaries of categorical maps, makes the appraise. What may be more important is the quality itself of the used in building the map.

aerial photographs normal

high definition- high altitude

orthophotos, orthophomaps satellite images

Meteosat NOAA

LANDSAT LANDSAT TM

SPOT-XS/P JERS-1 ERS-1 (slar) IRS-1

1:30k-60k

1: 100k + 1: 30k- 100k

Swath

5 km

185x170 m

185x170 m 60km

75km 100km 150km

variable ***

variable ***

same of the topographic map Resolution/avera ge accuracy 7.5km

1.5/6km 120m 45m

10-20/15-30m

18m/25m 30m/45m 36-72m/50-110m

issue of positional accuracy difficult to sources and the consistency and methodology

low 5 - 10 years medium 2 - 5 years low to medium

30min

12hrs 17 days 17 days 28 days 44 days

22 days

National/ Federal Mapping Authorities.

Military forces

Meteo. Depts.

Meteo. Depts.

Mapping, natural resources, environmental, authorities, Remote sensing centres

*** The scale within a photographic image is not constant. The positional accuracy depends mainiy on flight height, ground relief and the inclination of the camera axis.

- 17-

A framework for the establishment of national GIS

Section Two; GEOGHAPHIC INFORMATION MANGEMENT

NON-GRAPHIC

studies, reports, publications, statistical data, socio-economic data, etc.

ECA/NRD/CRSU/93- 7 Preliminary document - Revision June 1994

Everybody

Further characteristics of geographic data and information sources

The topographic map is probably the most important data source. It is the regular representation of earth features and phenomena, normally constructed following rigorous procedures and specifications, and constitutes the standard reference upon which other databases are build. However, (i) Offers

"limited" environmental information: while general areas of forest, rangeland, deserts, cultivated and populated land are shown, it doesn't provide major detail of the characteristics within those areas, unless for very large scales; (ii) it is costly and time consuming to establish and maintain, hence, it is usually out-of-date. Even considering the "ideal scheme" whereby the national mapping can be kept up within a 5 and a 10 year revision cycle for dynamic and less dynamic areas respectively, rapid changes, whether gradual of sharp, will not be detected timely.

Concerning thematic maps, with some exceptions, these are not produced and revised in developing countries in a systematic manner, and in many cases only offer a partial coverage, and again, if they are not up-to-date their utility will be reduced or disappear.

Cadastre and land tenure maps, if produced in the context of a multi-disciplinary system, are invaluable source of environmental information. These maps will not only portray the boundaries of the parcels, but also how the land is possessed and used. A sound cadastre system will equally indicate the type of soils and their aptitude or vocation. Maps depicting zones of homogenous characteristics, such as slope, erosion, road accessibility, water availability, value of the land, etc, are natural by-products of such systems.

Aerial photographs have been for many years the basic and most appreciated source of data for studies of earth related phenomena, and still constitutes the major input for topographic and thematic mapping. It offers a richness of detail unsurpassed by any other source, and can be obtained in standard black and white (B&W), in natural colours, in B&W infrared -where the infrared reflection of the earth surface is captured-, or in (false) colour infrared, providing a vast range of possibilities for the detection and interpretation of the imaged elements. However, each photography has a reduced coverage and for some applications, such as geomorphological studies, and when the territory to be analyzed is vast, it will only give a limited vision, and often too many photographs will have to be treated. It has the inconvenience that it requires of cloud free weather and therefore the obtention of a good aerial coverage becomes very difficult in tropical areas, sometimes impossible. The cost (around $4/km2 for small scales) is also a limiting factor to embark in periodic and systematic aerial photo coverage. The recent development and production of aerial cameras with forward image motion compensation (FIMC), have marked one of the important revolutions in mapping technology in the last decade. They have allowed the use of high resolution film with which we obtain images with the same detail and accuracy as those from photographs having half the scale, being specially suitable

-18-

A framework for the establishment of national GIS ECA/NRD/CRSU/93-7 Section Two: GEOGRAPHIC INFORMATION MANGEMENT Preliminary document - Revision June 1994

for high altitude photography. Scales of 1:100.000 and above can now be used. This advancement allows speeding up of production while cutting down costs notably.

Radar imagery: From active sensors installed in airborne or space platforms, radar images of the earth surface are obtained independently of weather or light conditions, through cloud coverage and

at any time of the day or night. However, major inconveniences of airborne radar images are the heterogenous scale within the image, and the "shadows" in areas of high relief caused by ground slopes. Although hi these images the geomorphology and the water areas show very well, their interpretative capacities for land use and land cover still have to be proved.

Space technology has brought a new dimension to resource mapping. Satellite images and then- corresponding orthophotos or spacemaps, provide general views of any area of the globe, at low cost and at regular intervals hi time with a high repetitivity. It has pros and cons, detractors and defenders, but it is proving to be the best option so far available that can effectively solve the major land and environmental information gaps of developing countries, including the completion and revision of their national topographic mapping. In Asia and America they are being introduced and accepted as standard operations in their cartographic and natural resources institutions and programmes. In this regard, it is gratifying to note that some countries in the region have already incorporated the use of satellite images in their regular programmes ; others are in the verge of doing so.

It has to be said that in some regions of Africa, cloud coverage imposes serious restrictions to obtain good satellite unages, but these restrictions are even more serious for conventional aerial photography.

In such areas, the high repeatability of the satellite image is a factor favoring the possibility of obtaining workable images with reasonable delays. In this respect, the radar capabilities of ERS-1 satellite should prove of particular interest. The high resolving power and other characteristics of the system make it very well-suited for territorial applications in the tropical belt in Africa because it is not affected by meteorological problems. Radar, moreover, has a special sensitivity for the morphological characteristics of the surface and the topography, and could prove to be a particularly effective tool in the production of forest and geological maps.

From the rigorous point of view of planimetric accuracy, LANDSAT TM imagery proves to be perfectly apt for the revision and up-dating of maps at scales up to 1:150.000. Strictly speaking, it should be the system selected with priority over other possible systems, both satellite and conventional. Furthermore, for vast areas of the African continent, where the landscape is less developed, a more relaxed accuracy can easily be accepted and the use of LANDSAT imagery be extended without major restrictions or hesitations to the scale of 1:100.000.

Similarly, SPOT imagery should be the system selected with priority over conventional systems for planimetric mapping and revision at scales ranging from 1:150.000 up to 1:66.000, and for remote and less developed areas, up to 1:50.000 (SPOT P).

Of all the operational satellites, SPOT is the only one offering altimetric determination capabilities.

The standard accuracy attained has been assessed at 5 to 10+ meters, depending on the conditions (identification of control points, temporal separation and the B/H ratio). In general, the altimetric

accuracy is recognized to be compatible with topographic mapping with minimum contour intervals

-19-

A framework for the establishment of national GIS ECA/NRD/CRSU/93-7 Section Two: GEOGHAPHIC INFORMATION MANGEMENT Preliminary document - Revision June 1994

of about 30 to 40 meters.

Precision in positioning is not the only factor that determines the applicability of satellite imagery in mapping. Another equally important factor is the capacity to interpret from the image the elements of interest that must be consigned in the map. Such capacity is directly related to the spacial resolution or pixel size of the image, which conditions the level of detail detection and identification.

It is beyond any doubt the factor of major controversy regarding satellite mapping suitability at medium and small scales, ranging from those who totally ostracize it and those who accept it.

Tables 2, 3 and 4 in Annex 1 shows us how the conditions of accuracy and interpretability are ordinarily met with LANDSAT TM and SPOT satellite images.

3. Data organization

When studying one particular environmental or development problem, the expert has to identify first what type of data is needed to solve his problem. Then he has to locate data sources, find it is existent and available and if its quality is OK. This location and assessment of data may be, as it is often, a tall order. The expert may not be able to identify all needed sources, and much less the quality of the data. It may also happen that, initially, he can not identify all his needs, regardless how keen or specialist he is in the matter, as there may be valuable documents that can assist in solving the problem but of which I am not aware of.

If the documented catalogue mentioned at the beginning of this presentation is available and properly maintained and accessed, then the puzzle is easily solved. If it is not, the expert can be in deep trouble. Indeed, the data and information required for any environmental aspect is a set constituted by numerous subsets of data pertaining to different sectors. On the one hand, modern technological development is leading to a convergence of what previously were discrete disciplines. On the other, we can not handle any more one resource or one environmental aspect in isolation from others resources or aspects, as it was done in simpler times. In addition to sound spatial and temporal data, information and knowledge on discrete aspects of the environment, which eventually can be obtained independently for each of them, their actual understanding and management cannot be properly conducted without considering their interaction. As an example, water resources is dependent on climate, rain fall, soil structure, geomorphology, vegetation cover, weather behavior, hydraulic and hydrological infrastructure, etc. Food security will be influenced by land use, land potentiality, natural vegetation, techniques in use, human influence, industrial development, population growth and migrations, industrial development, and many others.

The identification of a suitable location for a refuge camp, will require, e.g. to locate land with water accessibility, benevolent climate conditions, a particular soil structure, adequate slopes, proper drainage, etc, which are all find in different documents of different quality.

In conclusion, the data and information set of any aspect of development is constituted by numerous subsets of data and information pertaining to different sectors.

-20-

A framework for the establishment of national GIS

Section Two: GEOGHAPHIC INFORMATION MANGEMENT

ECA/NRE^CRSU/93-7 Preliminary document - Revision June 1994

climate soil structure

rain-fall

geomorphology vegetation cover weather behavior hydraulic and

etc

hydrological infrastructure

etc

trade hard currency avail.

exports vs. imports

land use land potentiality natural vegetation techniques in use human influence industrial development population growth and land tenure policies migrations and many others.

the vegetation set the food security set

In a nutshell, the task of the planner is to identify and access the required subsets of information that are relevant to solve his problem, and discard redundant and inadequate data. Then, what is relevant has to be simplified, and synthesized, adding intelligence to it so that appropriate decisions are made.

In other words, two are the main issues to resolve: firstly, the provision of means for accessing accurate, reliable and timely data and information; and secondly, the provision of means for the integrated analysis of the different types of data that is made available, comprising both internal and external factors.

How can we achieve the above objective?

The current analog mode of the documents containing the spatial and non-spatial data and information, that is drawings and text on paper form, does not facilitate the extraction of the selected elements, neither the integration with other analog data, much less their analysis. First, it is not easy to obtain copies of the documents, and to physically manipulate them. Further, the data contained in analog form is of rigid nature, making it difficult to separate what is relevant from the rest of the information that is disturbing and adds confusion to the work. But what is more problematic are the difficulties inherent in integrating heterogenous data coming from an different sources, with variable resolutions, scales and inaccuracies. We can use projectors to blow up (or reduce) and plot a map at a different scale or we can use photographic enlargers. We can draw a series of overlays, one for each selected element, trying to overcome the geometric inconsistencies by visual interpolation, or forcing the features to fit the geometry of a particular map. Then we can, more or less visually, interpret the results and make synthesis-maps portraying those results. You will agree with me that when the amount of data sources is high and the area subject to analysis covers numerous map sheets, then the task of extracting and integrating the data is simply formidable. Another inconvenient of this form of presentation is that normally the results, including resulting maps, are normally consigned in a report, and the new maps are lost for other potential users.

Fortunately, today we have powerful tools hat have been perfected to do such a task. These are the Geographic Information Systems (GIS).

-21-